INDEX

Explanations

symbols and punctuation marks

oai_token-act-pair · gpt-3.5-turbo

occurrences of a special character sequence that appears to signify importance or emphasis

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 shroud

-0.77

 board

-0.69

 raft

-0.68

 sleeper

-0.67

 flying

-0.67

bun

-0.66

pei

-0.66

 guiActiveUnfocused

-0.66

 touches

-0.65

 dividing

-0.65

POSITIVE LOGITS

ł

1.18

ONSORED

0.99

¹

0.97

ª

0.94

Ĳ

0.92

ĸļ

0.91

£

0.90

idav

0.88

ĳ

0.87

sure

0.86

Activations Density 0.096%

symbols and punctuation marks

occurrences of a special character sequence that appears to signify importance or emphasis

No Comments

No Known Activations